|Publication number||US4780890 A|
|Application number||US 06/913,050|
|Publication date||Oct 25, 1988|
|Filing date||Sep 29, 1986|
|Priority date||Sep 29, 1986|
|Also published as||EP0263377A2, EP0263377A3|
|Publication number||06913050, 913050, US 4780890 A, US 4780890A, US-A-4780890, US4780890 A, US4780890A|
|Inventors||Michael G. Kane|
|Original Assignee||Microwave Semiconductor Corp.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (4), Referenced by (6), Classifications (9), Legal Events (5)|
|External Links: USPTO, USPTO Assignment, Espacenet|
This invention relates to digital timing circuitry, and it relates, more particularly, to a synchronization circuit which operates while the digital system clock is running to skip a clock pulse to achieve synchronization.
In time-division multiplexed digital communication systems, an operation known as demultiplexing requires synchronization in order to divide a serial data stream properly into a number of channels. Various techniques have been devised to achieve and maintain synchronization through out digital communication systems. For example, one technique called pulse stuffing is often used to insure that synchronization is maintained by providing an appropriate signal format in serial transmission systems. Another somewhat conversely related technique involves achieving synchronization by removing clock pulses from a series of clock pulses in response to a command or control signal. This technique is useful in synchronous digital systems wherein a machine cycle is composed of multiple clock cycles. All parts of the system must begin their machine cycles on the same cycle in order for the system to function synchronously. A circuit known as a pulse swallower is able to achieve synchronization for one part of a system while the system is running by swallowing clock pulses entering that part of the system until synchronization occurs.
For a variety of reasons, a trend has developed wherein digital systems operate at increasingly faster and higher rates. Conventional pulse swallowers are limited due to their circuit configurations to a maximum operating frequency of one over six τ (1/6 τ) where τ corresponds to the average delay interval associated with the operation of a typically loaded gate. In terms of the duration of a clock cycle, the period of the clock cycle must be no shorter than six τ. Although integrated circuit technologies have developed with less delay to reduce the value of τ, the previous frequency constraint of conventional pulse swallowers still presents a degradation in the maximum frequency of operation of digital communication systems.
It is an object of the present invention to provide a synchronization circuit configuration capable of reliable operation at clock frequencies up to one over four τ (1/4τ).
It is a related object to provide a pulse swallower utilizing a local feedback loop as a speed enhancement technique.
In an illustrative embodiment of a pulse swallower in accordance with the invention, logical means is provided for enabling a gate in a signal path for clock pulses and for disrupting the signal path by disabling the gate for a predetermined interval in response to a control signal which eliminates only one clock pulse at an input clock frequency corresponding to the reciprocal of four gate delay intervals where each gate delay interval is the average loaded gate delay of the gates used in the embodiment.
In some of the aspects of the invention, logical circuitry receives the control signal and produces the enable signal which extinguishes itself four gate intervals after the end of a pulse of the input clock signal. The logical circuitry includes two bistable circuits; one bistable circuit has an input for recieving the control signal while the other has an input for receiving the input clock signal. Each bistable circuit has a data input and one data output. The data input of the first bistable circuit has a fixed high level signal applied thereto and its one data output is applied to the data input of the second bistable circuit. The one data output of the second bistable circuit provides the enabling signal for the gate in the signal path.
In some of the further aspects of the invention, the second bistable circuit has a complementary data output of the one data output. The complementary data output is connected to another data input of second bistable circuit to serve a local feedback loop for extinguishing the enable signal in its high state after one clock cycle. The first bistable circuit is a D-type flip-flop having a reset input. The one data output of the second bistable circuit is applied to the reset input. The signal path includes a serially connected inverter and a NOR gate which serves as the gate responsive to the control signal input for selectively eliminating single clock pulses.
Features of the invention and additional objects of the invention will be more readily appreciated and better understood by reference to the following detailed description which should be considered in conjunction with the drawing.
FIG. 1 is a circuit diagram of a conventional pulse swallower.
FIG. 2 is an internal diagram of flip-flop suitable for the pulse swallower of FIG. 1.
FIG. 3 is a waveform diagram illustrating faulty operation of the pulse swallower of FIG. 1 under certain conditions.
FIG. 4 is a circuit diagram of a pulse swallower in accordance with the invention.
FIG. 5 is a gate level diagram of a suitable flip-flop advantageously employed in the output portion of the pulse swallower of FIG. 4.
FIG. 6 is a waveform diagram illustrating normal operation of the pulse swallower of FIG. 4 under the same conditions as those utilized in FIG. 3.
FIG. 1 provides a circuit diagram for a typical conventional pulse swallower. Basically, the circuit of FIG. 1 has a clock signal applied to input 11 and delivers an output clock signal at output 12. Generally, the output corresponds to the input except for the delay introduced by inverter 13 and NOR gate 14 which becomes more significant as the frequency of the clock signal is increased.
In response to a negative transition of the swallow signal applied to input 16, which is the clock input of flip-flop 17, the Q1 output goes high. Upon the occurrence of the next negative transition of the clock signal, the Q2 output of flip-flop 18 goes to a high state. At the time that Q2 goes high, the output of inverter 13 remains high while the output of NOR gate 14 is low. After Q2 goes high, the output of NOR gate 14 remains low even after the input clock signal goes high.
At this point in time, the clock swallowing cycle begins. In addition to forcing the output of NOR gate 14 low, the high level on Q2 resets flip-flop 17 via conductor 19; as a result the Q1 output of that flip-flop returns to a low level. The next negative transition of the input clock signal passes the Q1 low level on to the Q2 output of flip-flop 18. Once Q2 is at a low level, the output of NOR gate 14 can once again change in response to the changes of the clock signal at input 11. However, at the same time Q2 of flip-flop 18 returns to a low level, the input clock signal is low, the output of inverter 13 is high, and the output of NOR gate 14 remains low until the next positive edge of the clock input signal. Thus one entire clock pulse has been swallowed by the circuit response to a falling edge or transition of the swallow signal.
FIG. 2 provides a gate level internal diagram of D-type flip-flop 21 which is suitable for both flip-flop 17 and 18 in FIG. 1. Flip-flop 21 has complementary data inputs, D and DB, and complementary outputs are Q and QB. In addition, flip-flop 21 has a clock input and a reset input. Since such D-type flip-flops are well known to those skilled in the art, further discussion of its internal operation is not warranted. However, the overall operation of the pulse swallower of FIG. 1 at high clock frequencies is important to understanding the full significance of the present invention.
FIG. 3 is a waveform diagram illustrating operation of the pulse swallower of FIG. 1 at a clock frequency equal to one over five τ (1/5 τ). As expected, in response to negative edge of the swallow signal one loss clock pulse should be present at output 12. The clock signal input is at the top of FIG. 3 while the clock output appears at the bottom of FIG. 3. What is evident is that two clock pulses are swallowed. It should be pointed out that when clock pulses are not being swallowed, the speed of circuit operation is only limited by the bandwidth of serially connected inverter 13 and NOR gate 14.
During a clock swallowing cycle however, the speed of operation is limited by the time required for Q2 to go low for enabling NOR gate 14. This requires flip-flop 17 to be reset, and for the Q1 output of that flip-flop to be adequately established at the correct level as an input signal to D of flip-flop 18. The delay between the falling edge of the input clock and Q2 going high is three τ, where τ is the average loaded gate delay. The delay between Q2 going high and flip-flop 17 resetting is two τ. The set-up time of flip-flop 18 is also τ. Thus, the total amount of time that must be available during the clock swallowing cycle is six τ. As a result, the maximum frequency of operation s one over six τ (1/6τ). The addition of a conventional pulse swallower to the fastest counters available which are capable of operating at frequencies up to 1/4τ will reduce the frequency of operation by about one third or 33%.
In FIG. 3, the complementary outputs of both flip-flops 17 and 18 are shown. In addition, waveform A is the output of inverter 13. The signal is an inversion of the clock input signal delayed by the response time of inverter 13. The output clock signal available at terminal 12 is again inverted by NOR gate 13 as well as slightly delayed.
FIG. 4 is a diagram of pulse swallower 41 in accordance with the principles of the present invention. As one would expect, pulse swallower 41 has equivalent inputs and an equivalent output to that of the pulse swallower of FIG. 1. In fact, inputs 42, 43 and output 44 including inverter 46, NOR gate 47 and flip-flop 48 are identical to the corresponding elements in FIG. 1. The primary differences between the pulse swallower of FIG. 4 from that of FIG. 1 lie in the way flip-flop 49 is connected and its internal circuitry which is depicted in FIG. 5. These differences enable the pulse swallower of FIG. 4 to operate correctly at frequencies of up to one over four τ (1/4τ). Accordingly when this pulse swallower is used in conjunction with a high speed counter, no degradation or reduction is experienced in the maximum operating frequency of the overall circuit.
As is evident from the circuit of FIG. 4, a local feedback loop via conductors 51 and 52 around flip-flop 49 forces this flip-flop to terminate the high state signal at its Q2 output after one cycle of the input clock signal. Therefore since this feedback loop does not include flip-flop 48, it is much faster than the equivalent feedback loop of FIG. 1 on conductor 19.
The delays associated with the local feedback loop on conductor 51 will now be described. As previously described in connection with the pulse swallower of FIG. 1, the delay between the falling edge of the input clock signal and the changing states of the signal at Q2 and QB2 is still three τ, and the set-up time of flip-flop 49 is τ.
FIG. 5 illustrates the internal circuitry of flip-flop 49. The local feedback of conductor 51 provides one input to AND gate 61 which is integrated with NR gate 62 which it feeds. In other words, AND gate 61 and NR gate 62 form a single composite AND-OR-INVERT gate. It should also be pointed out that the additional NOR gates in FIG. 5 which receive inputs (NOR gates 66-68) are cross coupled together with AND-OR-INVERT gate 61,62. This highly paralled arrangement provides concurrent operation of the associated gates for providing quick intermediate inputs to output NOR gates 71 and 72.
With the foregoing in mind, attention is directed to the overall pulse swallower of FIG. 4. It therefore becomes apparent that the total amount of time that must be available during the clock swallowing cycle is four τ. Thus, the maximum frequency of operation is one-over-four τ(1/4τ). As a result, the utilization of a pulse swallower according to FIG. 4 enables a high speed counter to operate at the highest speed which it is capable of operating.
FIG. 6 provides a timing waveform diagram for the pulse swallower of FIG. 4. This timing waveform diagram provides the digital signals of the pulse swallower of FIG. 4 which correspond to those of the pulse swallower of FIG. 1. However, the clock output signal line at the bottom of FIG. 6 clearly illustrates that only one pulse is swallowed as desired for correct operation. Although the operating frequency is one-over-five τ (1/5τ), this operation will continue up to frequencies of one-over-four τ (1/4τ).
There has thus been shown and described a novel pulse swallower which fulfills all the objects and advantages sought therefor. Many changes, modifications, variations and other uses and applications of the subject invention will, however, become apparent to those skilled in the art after considering this specification and the accompanying drawing which disclose the preferred embodiments thereof. All such changes, modifications, variations and other uses and applications which do not depart from the spirit and scope of the invention are deemed to be covered by the invention which is limited only by the claim which follow.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US3755748 *||Mar 6, 1972||Aug 28, 1973||Motorola Inc||Digital phase shifter/synchronizer and method of shifting|
|US4131854 *||Jul 15, 1977||Dec 26, 1978||Siemens Aktiengesellschaft||Switching circuit for regulating the repetition rate of clock pulses|
|US4191849 *||Oct 26, 1978||Mar 4, 1980||Gte Automatic Electric Laboratories Incorporated||Data synchronization circuit|
|US4578799 *||Oct 5, 1983||Mar 25, 1986||Codenoll Technology Corporation||Method and apparatus for recovering data and clock information from a self-clocking data stream|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5018168 *||Feb 9, 1990||May 21, 1991||Fujitsu Limited||Clock signal conversion circuit|
|US5192914 *||Jul 15, 1991||Mar 9, 1993||Fujitsu Limited||Clock control circuit for suppressing clock pulses|
|US5378934 *||Dec 16, 1992||Jan 3, 1995||Hitachi, Ltd.||Circuit having a master-and-slave and a by-pass|
|US5517521 *||Jun 25, 1993||May 14, 1996||Digital Wireless Corporation||Method and apparatus for synchronization between real-time sampled audio applications operating full-duplex over a half-duplex radio link|
|US6028425 *||Jan 13, 1997||Feb 22, 2000||Hewlett-Packard Company||Signal generator|
|US6316943 *||Dec 22, 1999||Nov 13, 2001||Agilent Technologies, Inc.||Signal generator|
|U.S. Classification||375/354, 370/505, 327/23, 327/176|
|International Classification||H04L7/00, H03K23/66, H04J3/06|
|Sep 29, 1986||AS||Assignment|
Owner name: MICROWAVE SEMICONDUCTOR CORPORATION, 100 SCHOOL HO
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNOR:KANE, MICHAEL G.;REEL/FRAME:004618/0880
Effective date: 19860925
|Dec 5, 1988||AS||Assignment|
Owner name: SIEMENS AKTIENGESELLSCHAFT, BERLIN AND MUNICH, A C
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNOR:MICROWAVE SEMICONDUCTOR CORPORATION;REEL/FRAME:005006/0706
Effective date: 19881115
|May 28, 1992||REMI||Maintenance fee reminder mailed|
|Oct 25, 1992||LAPS||Lapse for failure to pay maintenance fees|
|Jan 5, 1993||FP||Expired due to failure to pay maintenance fee|
Effective date: 19921025